Computer-aided prediction of new anti-HIV compounds, derived from substructures of 2-amino-6-arylsulfonylbenzonitriles and cyclotriazadisulfonamide analogues, has been reported. A ligand-based approach, namely MIA-QSAR, and a docking evaluation were used to model the title compounds, macrocycles containing a trisubstituted benzene moiety. According to the MIA-QSAR method, predicted potencies for proposed compounds were up to seven times higher than that of the experimentally most active compound of training set. Moreover, we have used docking approaches to study the binding orientations and predict binding affinities of these compounds in CD4 receptor.